AI Agent Security: New Threats and the Future Attack Surface

The series closes with the founder of the OWASP GenAI Security Project mapping the agentic AI threat landscape for healthcare. AI agents are moving from proof-of-concept to production, managing scheduling, clinical documentation, revenue cycle workflows, and prior authorizations, and they're bringing an attack surface most security teams aren't prepared for. Steve Wilson points to recent attacks that show what happens when a legitimate platform is turned into a weapon, then asks us to imagine that same dynamic with AI agents: autonomous systems holding elevated permissions across clinical environments. As he puts it, prompt injection, agent hijacking, and cascading failures are documented vulnerabilities today, not theory. Featuring: Steve Wilson — Chief Product Officer, Exabeam; Founder & Co-Chair, OWASP GenAI Security Project Moderator: Ed Gaudet — CEO & Founder, Censinet